ML-MIX documentation

Welcome to the documentation for ML-MIX, a LAMMPS plugin for performing efficient spatial mixing of different pair_styles!

This documentation covers:

• Setup and installation of the LAMMPS plugin and Python packages required for running the provided case studies.

• An overview of the plugin.

• Simple examples demonstrating how to use the plugin.

• Detailed LAMMPS-style documentation on the commands provided by the plugin.

• Documentation on the constrained fitting of small, fast linear ACE and UF3 potentials.

Contents:

• Setup
  • Cloning this Repository
  • Cloning LAMMPS
• Building ML-MIX
  • In-Source Build (Required for KOKKOS)
  • Plugin Build (Without KOKKOS)
  • Loading the ML-MIX Plugin
  • Installing Python Packages
  • Installing ACEpotentials.jl for ACE fitting
  • Installing UF3 for UF3 fitting
• Package Overview
  • Command overview
  • Tested Compatible ``pair_style``s
  • CPU pair_styles
  • GPU pair_styles
  • Notes and Contributions
  • Why Use This Package?
  • How it works: Data flow
  • How it works: Region building
  • Time Decay Hysteresis
  • How it works: Force mixing
  • Mitigating energy drift
• Examples
  • Basic Examples
  • Advanced Examples
  • How to Run the Examples
• LAMMPS Style Documentation
  • fix mlml
  • pair_style hybrid/overlay/mlml
  • fix langevin/mlml command
• Potential fitting
  • Constrained Potential Fitting
  • Fitting Process